Click here to


Are you sure ?

Yes, do it No, cancel

Image-Based Correction of External Respiratory Surrogate Signal for Accurate T2-Weighted 4D MRI

J Uh1*, M Kadbi2 , C Hua1 , (1) St Jude Children's Research Hospital, Memphis, TN, (2) Philips Healthcare, Gainesville, FL


(Wednesday, 8/1/2018) 1:45 PM - 3:45 PM

Room: Karl Dean Ballroom B1

Purpose: To improve the image quality of slice-wise retrospective-sorting, transverse T2-weighted 4D MRI which provides favorable image orientation and contrast for radiation treatment planning, but suffers from incorrect external respiratory surrogate signal particularly with irregular breathing.

Methods: Serial volumes of T2-weighted images were acquired on 4 volunteers using a single-shot turbo spin-echo sequence (sampling rate, 1.1-2.4 Hz; scan time, 7.6-9.7 minutes), while a pneumatic sensor fastened on the upper abdomen recorded external respiratory signal. The images acquired at the same slice location were processed by principal component analysis (PCA) to derive indices representing internal respiratory motion. The external surrogate signals were adjusted using the PCA indices to produce corrected signals. The impact of the correction on reconstructed 4D MRI was studied in relation with breathing irregularities scored by the coefficient of variation (CV) of peak-to-peak amplitude. The coherence among slices within a volume was quantified by the mean of structural similarity index (mSSIM) between adjacent slices. The proposed 4D MRI was qualitatively compared to a prospective-sorting 4D MRI that used internal navigator.

Results: Visual inspection showed that the sorting artifacts were reduced in all 4D images by using the corrected surrogate signals. The mSSIM significantly increased (by 2.2% and 4.1%) for the subjects with greater breathing irregularities (CV, 26% and 25%), while the impact was minimal (mSSIM increase, 0.2% and 0.8%) for those showing relatively regular breathing (CV, 13% and 16%). The image qualities were comparable to the prospective-sorting 4D MRI of the similar scan time.

Conclusion: The image-based correction of external respiratory surrogate signal improved the retrospective-sorting T2-weighted 4D MRI, for which the use of only internal surrogates was not feasible because of the interference with navigator and slow sampling rate. The proposed method provides an alternative to the prospective-sorting 4D MRI especially for patients with irregular breathing pattern.

Funding Support, Disclosures, and Conflict of Interest: This work has been supported in part by Philips Healthcare and ALSAC.


MR, Treatment Planning, Organ Motion


IM/TH- MRI in Radiation Therapy: MRI for treatment planning

Contact Email